Title :
Minimum Classification Error for Large Scale Speech Recognition Tasks using Weighted Finite State Transducers
Author :
McDermott, Erik ; Katagiri, Shigeru
Author_Institution :
NTT Commun. Sci. Labs., NTT Corp., Kyoto, Japan
fDate :
March 18-23, 2005
Keywords :
decision trees; learning (artificial intelligence); minimisation; signal classification; speech recognition; MCE training; acoustic models; audio training set; discriminative training; large vocabulary speech recognition; lattice-derived WFST; lecture speech transcription; maximum likelihood estimation; minimum classification error method; phonetic decision trees; telephone-based name recognition; triphone models; weighted finite state transducers; Acoustic measurements; Acoustic noise; Acoustic transducers; Laboratories; Large-scale systems; Lattices; Maximum likelihood estimation; Speech recognition; Target recognition; Vocabulary;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP '05). IEEE International Conference on
Print_ISBN :
0-7803-8874-7
DOI :
10.1109/ICASSP.2005.1415063
